Impact of Galactic non-Gaussian foregrounds on CMB lensing measurements

TL;DR

This paper assesses how non-Gaussian Galactic dust foregrounds affect CMB lensing measurements, finding minimal impact for ACT but potential concerns for the Simons Observatory, highlighting the need for further dust bias characterization.

Contribution

It provides a comprehensive analysis of Galactic dust contamination on CMB lensing, using multiple models and estimators, and evaluates the impact for current and future experiments.

Findings

Galactic foregrounds cause less than 0.3σ bias for ACT.

Potential significant bias for the Simons Observatory under some models.

Highlights the importance of dust bias mitigation for large sky fractions.

Abstract

Weak gravitational lensing of the CMB has been established as a robust and powerful observable for precision cosmology. However, the impact of Galactic foregrounds, which has been studied less extensively than many other potential systematics, could in principle pose a problem for CMB lensing measurements. These foregrounds are inherently non-Gaussian and hence might mimic the characteristic signal that lensing estimators are designed to measure. We present an analysis that quantifies the level of contamination from Galactic dust in lensing measurements, focusing particularly on measurements with the Atacama Cosmology Telescope and the Simons Observatory. We employ a whole suite of foreground models and study the contamination of lensing measurements with both individual frequency channels and multifrequency combinations. We test the sensitivity of different estimators to the level of foreground non-Gaussianity, and the dependence on sky fraction and multipole range used. We find that Galactic foregrounds do not present a problem for the Atacama Cosmology Telescope experiment (the bias in the inferred CMB lensing power spectrum amplitude remains below $0.3\sigma$). For Simons Observatory, not all foreground models remain below this threshold. Although our results are conservative upper limits, they suggest that further work on characterizing dust biases and determining the impact of mitigation methods is well motivated, especially for the largest sky fractions.